A Novel Approach for Spoken Language Identification and Performance Comparison Using Machine Learning-Based Classifiers and Neural Network
| Autor: | Manthan Manavadaria, Vishal P. Tank, Krupal Dudhat |
|---|---|
| Rok vydání: | 2021 |
| Předmět: |
Artificial neural network
Computer science Speech recognition Gaussian Linear discriminant analysis Tree (data structure) Naive Bayes classifier symbols.namesake ComputingMethodologies_PATTERNRECOGNITION Classifier (linguistics) symbols Mel-frequency cepstrum MATLAB computer computer.programming_language |
| Zdroj: | Advances in Intelligent Systems and Computing ISBN: 9789811621222 |
| DOI: | 10.1007/978-981-16-2123-9_42 |
| Popis: | Spoken Language Identification (SLI) is the process of capturing a type of language of a speaker. In this research paper, the used database is created in three different languages, Gujarati, Hindi, and English. Language classification is performed using features like MFCC (Mel Frequency Cepstral Coefficients), Pitch, and average energy. Accuracy values of the created database are evaluated and compared using various pattern classifiers, namely, Fine Tree, Linear Discriminant, Gaussian Naive Bayes, Linear SVM, Fine KNN, and feed-forward neural network in MATLAB 2019. Performance using individual speech features and hybrid features are compared. Training time of all the classifiers is also evaluated to decide the best among all classifiers. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|